Hi,
does anyone know if the fix nvt (Nose-Hoover thermostat) implemented
in LAMMPS couple one thermostat to the entire system or if it is one
thermostat to each degree of freedom of each particle?
The one thermostat (or more) per degree of freedom is also know as
massive Nose-Hoover thermostat, it is useful for highly harmonic
systems.
Rodrigo Freitas
If is one thermostat coupled to the entire
system. Although there is a chains option
with fix nvt to use a handful of thermostats.
If you want to couple a thermostat to
each particle, see the fix langevin
command.
I think it makes little sense to use
Nose/Hoover on a per-particle basis.
Steve
If is one thermostat coupled to the entire
system. Although there is a chains option
with fix nvt to use a handful of thermostats.If you want to couple a thermostat to
each particle, see the fix langevin
command.I think it makes little sense to use
Nose/Hoover on a per-particle basis.
i think it does make sense, in fact, like it was
stated, there are even setups, where you
want to couple each degree of freedom
individually. this i what people use in gas
phase CPMD calculations and other
setups (e.g. path-integral simulations),
where it is difficult to get equipartitioning
of the kinetic energy.
unfortunately, langevin dynamics doesn't
have the proper canonical sampling. this
is what the CSVR thermostat algorithm
by giovanni bussi is supposed to cure.
one big problem of a massive NH-chain
thermostat in classical MD is it relative
"stiffness" requiring a shorter time step
and the high cost to propagate
NH-chains for 3N-3 DOFs.
cheers,
axel.